Energy intake is the most difficult part of the energy balance equation to measure. Using an in patient computerized vending machine system, food intake has been measured over 180 individuals. In individuals undergoing the study more than once the intra-class correlation coefficient is very high (r=0.9) indicating that these studies although performed in an inpatient setting are very reproducible. Eating during nighttime hours is a behavior that may predispose individuals to obesity. In our vending machine studies, we found that nearly 30% of volunteers ate from our vending machines during the night. These individuals consumed overall more calories and gained more weight over a 3 year follow-up period. In addition, these individuals have higher respiratory quotients measured in our respiratory chamber while on a weight maintenance diet indicating an increased drive for food intake. Future efforts will examine whether there are differences in anorexigenic or orexigenic hormones between night eaters and non-night eaters. Higher respiratory quotient, the ratio of carbohydrate oxidation to lipid oxidation, is known to predict weight gain. The mechanism of this increase in weight was thought to be secondary to a lower lipid oxidation in these individuals. Higher respiratory quotient measured during weight maintenance also predicted food intake during the ad-libitum vending machine study. However the component of respiratory quotient which best predicted food intake was higher carbohydrate oxidation indicating a role for carbohydrate (likely glycogen) balance in the regulation of short term food intake. To further investigate whether carbohydrate balance influences food intake, we fed individuals (in random order) a high fat versus high carbohydrate diet followed by ad libitum food intake. In preliminary analysis of 20 individuals, we did not find any difference in food consumption after either diet. Predictors of food intake in the past have demonstrated that individuals with lower fasting ghrelin concentrations (a hormone which is associated with increased food intake) actually ate more. However, follow-up studies on a larger group of individuals did not show any association between total ghrelin concentrations and food intake. Furthermore, total ghrelin concentrations did not decrease with overfeeding, and change in ghrelin concentration did not predict subsequent food intake. This data indicates that ghrelin may not have a role in control of day to day food intake, nor is it suppressed by overeating. In order to evaluate this further, another study arm is evaluating the effect of both total and active ghrelin concentrations following a period of fixed 3 day period overfeeding compared to a period of weight maintenance diet. Physical activity, energy expenditure, and basal body temperature are also evaluated during both overfeeding and fasting, and ad libitum food intake is evaluated following each diet period. Preliminary evaluation of this group indicates that individuals do not down regulate food intake following a period of overfeeding. Offspring born to mothers with diabetes (ODM) tend to have more rapid weight gain than those born to women without diabetes during the pregnancy (NON). Energy expenditure is not lower in ODM compared to NON. We also evaluated ad libitum food intake in ODMs versus NON. We found no difference in food intake. However there was a strong association with maternal food intake indicating this was the dominant effect.